Generally, a measuring device such as an electric power sensor or temperature sensor performs measurement periodically and outputs data indicating the measurement result.
The data structure of the data outputted from the measuring device is constituted by “date and time” at which the measurement was conducted, a “device ID” which can identify the corresponding measuring device uniquely, and a “measurement value” obtained by the measurement.
A method which prevails conventionally most widely for managing the data in the table of the database is the one that stores the data in the table without changing the original data format described above.
If all data of the same time are to be extracted for the use of, for example, monitoring, at once from the database which has been accumulated in the above format, since the time added to the data itself differs depending on the measuring device, it is difficult to compare the data which are measured at the same time.
According to the conventional management scheme, the data of the respective devices are stored all in the direction of column for each time. When, for example, the correlation between the temperature and electric power is to be obtained, a complicated process is required.
As opposed to the above scheme, for the object of comparing sensor data which are measured periodically, in a time series manner, Patent Literature 1 discloses a technique for generating a record segmented on the time base, and storing data in a record whose time is close to the time of the data, so that a plurality of pieces of data having the same time or times close to each other are stored in one record.